Even at first glance, there is something about the body on the street. The white paper under it is very clean, and the officers movements are completely free of purpose. “We need to remove the street,” one of them says with the firm hand, although her lips do not move. It is Amnesty International, well. But here is Kicker: It did not include any dialogue.
VEO 3 added, the new Google video generation model, this line on its own. Over the past 24 hours, I created dozens of clips depicting news reports, disasters and cartoon cats with the masked voice – some invented the model on their own. It is more than a little creeping and more sophisticated than I imagined. Although I do not think it will lead us to the wrong Day of the Resurrection so far, VEO 3 strikes me as an absolute machine of artificial intelligence.
Google Veo 3 foot In I/O this week, highlights its most important new ability: generating sound to go with your AI’s video. “We are entering a new era of creation,” explained by Google Gemini, Josh Woodward, in the keyword, describing it as “incredible realistic.” I had not completely sold, but after that, a few days later, I created a video of the news broadcaster announcing a fire in Edle. All it takes was a basic text directed, a few minutes, and a very expensive subscription Google Ai Ultra Plan. And you know what? Woodward was not exaggerated. It is realistic like hell.
I tried the news of the news anchor after seeing what Alejandra Carrabalo, a clinical coach at the Cyberlo Clinic at Harvard Law College, was able to produce it. One of its clips A news broadcaster announces the death of US Defense Secretary Beit Higseth. He did not die, but the clip is incredibly convincing. Published including a series of videos with Persons created from artificial intelligence protesting the claims used to create them He has 50,000 UPVotes on Redit. Scenes include disasters, a woman in a hospital bed using a breathing tube, and a character who is threatened at gunpoint – all with spoken dialogue and realistic background sounds. Real things!
I may be naive, but after playing with Veo 3, I am not so completely I was also worried as I was at the beginning. For beginners, there are clear handrails in place. You cannot claim it to create a video of Biden stumbling and falling. You cannot have a news broadcaster announcing the assassination of the president, or even creating a video of the executive director of the technology company that wears the shirt and the series laughs while the dollar around him is raining. This is the beginning.
However, you can generate some disturbing shit. Without any smart solutions, the VEO3 prompted to create a video of space needle on fire. Starting with my own image of Jabal Renier, I created a video of it that broke out with smoke and lava. Besides a clip of the news anchor who announces the aforementioned catastrophe, I can see how you can drink some harm easily with this tool.
Here’s the best news: It doesn’t look like a ready Deepfake machine. I gave her a few pictures of myself and asked for this to create a video with a specific dialogue and will not comply. I also asked to bring a pair of giant shoes in a picture to life and make them come out of the scene; I managed to stir the shoe across the sidewalk with some comic noise in the background.
I had an easier time to create videos when my demands were less specific, which I confirmed something My colleague Andrew Marino indicatedVeo 3 is excellent in creating a type of less inclusive YouTube content on YouTube targeting children.
If you are never exposed to an endless hole of garbage on YouTube Kids, allow me to get you. Imagine watching the worst three -dimensional display of a monster truck driving down the slope, and fell at the value -added tax of colored paint. Next to it, another monster truck is pushing to the bottom of another slope at another value -added tax – this time, a different color. Now see it again. And again. And again. There are hours of these things on YouTube designed to break up young children. These videos are usually not harmful, only the empty calories designed to lift the views that make cocomylone look Citizen Kane. In about 10 minutes with Veo 3, together a clip followed the same basic formula – complete with the Jaunty background music. But the most disturbing section for me is cartoon cats on a berth.
I thought it would be funny to complain to each other that the fish did not bite. Within just a few minutes, I had a full clip with two cats and some dialogue that was created from artificial intelligence I never wrote. If it is easy to create a 10 -second clip, then extending it to a seven -minute video clip will be trivial. In its current form, the clips return to the Veo 2 when trying to extend it to longer scenes, which remove the sound. But the way Google pushes these tools unabated forward, I can’t imagine that it will be long before you can edit a full video clip with Veo 3.
Frankly, I wonder if this type of use of the video created from artificial intelligence is a feature and not a mistake. Google showed us some video of artificial intelligence from real movie makers, Including Elisa McKenetWho works with Darren Arenovsky in a new movie with some of the elements that are created from artificial intelligence. Certainly, the artificial intelligence video can be an interesting tool in the right hands. But I think what we likely see is the spread of a kind of Nice pictures that artificial intelligence is very good in generating – This time, in Stereo.
